RESEARCH
Funding agencies in low- and middle-income countries: support for knowledge translation
Aide à la transposition dans la pratique des connaissances par les agences de financement des pays à revenu faible ou moyen
Organismos de financiación en países de ingresos bajos y medios: apoyo a la traslación de conocimientos
Cynthia CorderoI; Rachel DelinoI; L JeyaseelanII; Mary Ann LansangI; Juan M LozanoIII; Shuba KumarIV; Socorro MorenoIII; Merle PietersenV; Jose QuirinoVI; Visanu ThamlikitkulVII; Vivian A WelchVIII; Jacqueline TetroeIX; Aleida ter KuileVIII; Ian D GrahamIX; Jeremy GrimshawIX; Vic NeufeldX; George WellsXI; Peter TugwellVIII,1
IDepartment of Clinical Epidemiology, College of Medicine, University of the Philippines, Manila, the Philippines
IIDepartment of Biostatistics, Christian Medical College, Vellore, India
IIIClinical Epidemiology and Biostatistics Unit, School of Medicine, Javeriana University, Bogota, Colombia
IVUnit for Evidence Based Medicine, Madras Medical College, Chennai, India
VMedical Research Council-South Africa, Cape Town, South Africa
VIGridec-Clinical Epidemiology Unit, Federal Universidade do Estado de Sao Paulo-UNIFESP, Brazil
VIIFaculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand
VIIICentre for Global Health, Institute of Population Health, University of Ottawa, Ottawa, ON, Canada
IXClinical Epidemiology, Ottawa Health Research Institute, Ottawa, ON, Canada
XFaculty of Health Sciences, McMaster University, Hamilton, ON, Canada
XICentre for Global Health, Institute for Population Health, University of Ottawa, Ottawa, ON, Canada
ABSTRACT
OBJECTIVE: The aim was to describe how selected health research funding agencies active in low- and middle-income countries promote the translation of their funded research into policy and practice.
METHODS: We conducted inductive analysis of semi-structured interviews with key informants from a purposive sample of 23 national and international funding agencies that fund health research in Brazil, Colombia, India, the Philippines, South Africa and Thailand. We also surveyed web sites.
FINDINGS: We found a commitment to knowledge translation in the mandate of 18 of 23 agencies. However, there was a lack of common terminology. Most of the activities were traditional efforts to disseminate to a broad audience, for example using web sites and publications. In addition, more than half (13 of 23) of the agencies encouraged linkage/exchange between researchers and potential users, and 6 of 23 agencies described "pull" activities to generate interest in research from decision-makers. One-third (9 of 23) of funding agencies described a mandate to enhance health equity through improving knowledge translation. Only 3 of 23 agencies were able to describe evaluation of knowledge translation activities. Furthermore, we found national funding agencies made greater knowledge translation efforts when compared to international agencies.
CONCLUSION: Funding agencies are engaged in a wide range of creative knowledge translation activities. They might consider their role as knowledge brokers, with an ability to promote research syntheses and a focus on health equity. There is an urgent need to evaluate the knowledge translation activities of funding agencies.
RÉSUMÉ
OBJECTIF: Décrire comment certaines agences, qui financent la recherche en santé dans des pays à revenu faible ou moyen, favorisent la transposition sous forme politique et pratique des recherches financées.
MÉTHODES: Nous avons réalisé une analyse inductive d'entretiens semi-structurés menés avec des informateurs clés d'un échantillon choisi à dessein de 23 agences nationales et internationales, qui financent des recherches en Afrique du Sud, au Brésil, en Colombie, en Inde, aux Philippines, et en Thaïlande. Nous avons également effectué une enquête sur des sites Internet.
RÉSULTATS: Nous avons relevé un engament à transposer les connaissances en pratique dans le mandat de 18 des 23 agences de l'échantillon. Cependant, la terminologie utilisée était peu homogène. La plupart des activités mentionnées sont des efforts classiques de diffusion auprès d'une large audience, par le biais par exemple de sites Internet ou de publications. En outre, plus de la moitié des agences (13 sur 23) encouragent les liens et les échanges entre chercheurs et utilisateurs potentiels et 6 agences sur 23 décrivent des activités de type « pull » pour intéresser les décideurs aux travaux de recherche. Un tiers des agences (9 sur 23) indiquent dans leur mandat vouloir améliorer l'équité en matière de santé par une meilleure transposition dans la pratique des connaissances. Seules 3 des 23 agences sont en mesure de mentionner une évaluation des activités de transposition en pratique des connaissances. Nous avons en outre constaté que les agences de financement nationales faisaient de plus grands efforts pour assurer cette transposition que les agences internationales.
CONCLUSION: Les agences de financement ont entrepris des activités très diverses de transposition en pratique des connaissances. Elles peuvent se considérer comme ayant un rôle de courtier en connaissances et comme ayant la capacité de promouvoir une synthèse des recherches et une convergence de l'attention sur l'équité en termes de santé. Il est urgent d'évaluer les activités de transposition en pratique des connaissances menées par les agences de financement.
RESUMEN
OBJETIVO: Describir cómo algunos organismos de financiación de investigaciones sanitarias que operan en países de ingresos bajos y medios promueven la traslación de las investigaciones que financian en políticas y prácticas.
MÉTODOS: Realizamos análisis inductivos de entrevistas semiestructuradas con informantes clave a partir de una muestra intencionada de 23 organismos nacionales e internacionales que financian investigaciones sanitarias en el Brasil, Colombia, la India, Filipinas, Sudáfrica y Tailandia. También sondeamos diversos sitios web.
RESULTADOS: Detectamos muestras de compromiso en favor de la traslación de conocimientos en el mandato de 18 de 23 organismos. Sin embargo, no había una terminología común. La mayoría de las actividades consistían en las iniciativas tradicionales de difusión de información entre un público amplio, por ejemplo a través de sitios web y publicaciones. Además, más de la mitad (13 de 23) de los organismos fomentaban el establecimiento de vínculos y el intercambio entre los investigadores y los usuarios potenciales, y 6 de los 23 organismos describieron actividades de «atracción» para generar interés por las investigaciones entre los decisores. La tercera parte (9 de 23) de los organismos de financiación tenían encomendado el fomento de la equidad sanitaria mediante la mejora de la traslación de conocimientos. Sólo 3 de los 23 organismos podían hacer una evaluación posterior de sus actividades de traslación de conocimientos. Además, observamos que los organismos de financiación nacionales hacían un mayor esfuerzo de traslación de conocimientos que los organismos internacionales.
ONCLUSIÓN: Los organismos de financiación participan en una amplia gama de actividades creativas de traslación de conocimientos y podrían tal vez estudiar su papel como intermediarios en ese ámbito, facultados para promover síntesis de investigaciones y un mayor énfasis en la equidad sanitaria. Es necesario evaluar urgentemente las actividades de traslación de conocimientos de los organismos de financiación.
Introduction
For knowledge to benefit society, it needs to be shared, communicated and translated into policy, practice or community action.1 Increased commitment to knowledge translation is reflected by the 58th World Health Assembly's declaration in 2005, which encouraged enhanced knowledge transfer.2 Several international initiatives focus on knowledge translation in low- and middle-income countries (LMICs) such as the Overseas Development Institute's RAPID programme (Research and Policy in Development), the WHO/PAHO EVIPNet initiative (Evidence-Informed Policy Networks) and the WHO Knowledge Management and Sharing initiative.
The WHO Department of Knowledge Management and Sharing defines knowledge translation as: "The synthesis, exchange and application of knowledge by relevant stakeholders to accelerate the benefits of global and local innovation in strengthening health systems and improving people's health."3 Because of the dearth of primary research performed in their own countries and the disproportionately low research resources available, LMICs need to engage in the translation of knowledge that is cost-effective and applicable to their local settings.4
Knowledge translation is a complex and nonlinear process, and is generally slow, particularly in LMICs.5,6 Slow knowledge transfer can result in inappropriate care. Many examples in LMICs have shown variations in practice despite established guidelines; for example, antibiotic prophylaxis with caesarean section,7 management of acute myocardial infarction8 and management of pneumonia.9 In one example, a study of Shanghai hospitals found that more than 70% of births involved clinical practices that are ineffective and should be avoided based on the best available evidence from the Cochrane Library.10
Knowledge translation may help bridge the knowdo gap, particularly in disadvantaged populations.3 Utilization of treatments with demonstrated effectiveness, such as immunization, oral rehydration for diarrhoea and treatment for acute respiratory infection, is up to 50% lower for the poorest.1113 Knowledge translation interventions that enhance access, diagnostic accuracy, provider compliance or consumer adherence could enhance community effectiveness of interventions in disadvantaged populations.14
Because research funding agencies are the gatekeepers to funds for conducting research, they may be able to encourage knowledge translation and exchange by their funding recipients. They can also actively disseminate information, involve end users in prioritizing research topics and fund implementation research. However, little is known about funding agency policies to promote knowledge translation.
This project was designed as an exploratory, descriptive study to increase understanding of the knowledge translation policies and activities of applied health research funders within LMICs and international funding agencies.
Methods
We conducted inductive analysis of semi-structured interviews with key informants from a judgement sample of funding agencies supplemented by document analysis from the agency web sites, including strategic plans, mandate and application procedures. This method provides a richness of data that cannot be assessed using questionnaire surveys since participants could respond freely as well as illustrate concepts with examples and the interviewer could probe for more details.15 Document analysis and findings from interviews were triangulated to present a complete picture of knowledge translation activities. We used the Lavis framework of push, pull, linkage/exchange and integrated efforts to classify knowledge translation activities.16
Sampling
We selected six LMICs, based on the presence of substantial within-country health research funding: Brazil, Colombia, India, the Philippines, South Africa and Thailand. None of these countries were among the least developed countries, where external funding agencies would be responsible for a larger proportion of health research funding (e.g. Bangladesh or Mozambique).17 Because this is an exploratory study of knowledge translation, we chose to use criterion-based purposive sampling, a non-probability sampling method that selects informants based on predefined criteria.18 As with other non-probability sampling methods, purposive sampling does not produce a sample that is representative of a larger population, but it is useful to study a clearly defined group. Our criterion for selecting funding agencies was the extent to which they funded applied health research. We selected a total of 14 national funding agencies from these six LMICs and nine international funding agencies, based on these criteria. Some country investigators applied additional criteria that are listed in Table 1. For each agency, we aimed to interview three key informants: someone from senior management with strategic responsibility, a research manager with responsibility for applied research programmes and a knowledge transfer officer. We interviewed key informants from 23 agencies between September 2003 and September 2004 (Table 1).
Interviews
The interviews were conducted face-to-face or via telephone by one of the authors, using a semi-structured interview framework (Table 2). Participants were asked to provide relevant documents or web sites that contained policy statements on knowledge translation as well as copies of grant application forms. Data was extracted using the same framework as the interview guide.
The interview guide was translated into Portuguese, Spanish and Thai. Each translation was back-translated into English by a second translator who had not seen the original English version. The English back-translation and the original were then compared. If the back-translated items and the original did not agree, the first translator conducted a second translation. A second back-translation was repeated. This process continued until the translation was judged satisfactory.
The audio-tapes were transcribed verbatim and verified by the interviewer before analysis. Transcripts were coded in their original language, and then translated to English to permit comparison of the findings from all the countries using the same approach used to translate the interview guide.
Two types of bias threaten this type of semi-structured interview and inductive analysis: description bias and interpretation bias. To minimize description bias, we transcribed interviews verbatim and used back-translation methods to ensure accurate translations. To minimize interpretation bias, we asked agency interviewees to verify data and we verified the coding with all co-investigators.
Analysis
We used inductive analysis to code and categorize data.19,20 We identified eight main themes: role of agency, background, researcher requirements, application process, dissemination activities, agency initiatives, evaluation and target audience. We further identified subcategories within each of these codes. Each of the LMIC investigators used these codes and subcategories to classify their data. The initial coding of all the data was performed by the interviewers in the LMIC and the co-investigator in that country.
To ensure that analysis was consistent between countries, we checked the classification of the verbatim transcripts at the central coordinating office in Ottawa, Canada, and finalized the coding by consensus through conference calls and e-mails with the investigators to ensure common understanding. We verified the final coding with the interviewees, allowing them to add or update information.
The analysis of this hypothesis-generating study focused on the nature of the knowledge translation activities of funding agencies and their perception about needs for improvement. We did not aim to compare funding agencies, hence individual results for each funding agency are not presented. Furthermore, because we did not interview all departments within each agency, we could not be certain that we had captured all knowledge translation activities.
Results
Coding
We developed the final coding of each interview by consensus discussion with the country teams and the Ottawa team. We kept records of the changes to the coding based on consensus discussion. We found that 89% of the coded text was identical between the original country team coding and the final coding. Most of the differences in coding were due to country teams placing descriptions of specific activities into the five general activities of the funding agency, which were intended to contain broad approaches rather than specific activities.
Analytical framework
Based on analysis of the interviews, we defined five broad categories of funding agency activities related to knowledge translation as follows: (1) "pull" was defined as: activities where the research agenda was set by policy-makers, activities that aimed to increase skills and capacity of policy-makers to use research evidence; (2) "push" was defined as: activities that encouraged researchers to communicate effectively with decision-makers; (3) "linkage/exchange" was defined as: creating linkages between researchers and policy-makers (e.g. workshops, conferences or knowledge brokers); (4) "communication" was defined as: the funding agency itself translating or communicating research results to research users and policy-makers; and (5) "funding opportunities" were defined as: specific funding opportunities that encouraged researchers to engage in knowledge translation strategies themselves.
We added the last two categories based on inductive analysis since communication efforts and funding opportunities were described as two important ways that funding agencies support knowledge translation. These categories did not fit into the Lavis framework of push, pull and linkage/exchange.
We found that these five codes for general knowledge translation activities were mutually exclusive, i.e. despite allowing double-coding of text where relevant, no text was placed in more than one of the five general activities. We found two cases from the 23 agencies where negotiation of meaning with the Ottawa team resulted in reclassifying push activities as pull activities.
Mandate
Thirteen of 23 agencies described a favourable political climate to knowledge translation, mainly due to increasing realization that research needs to infiltrate policy and action to benefit health. Respondents described the following barriers to knowledge translation: lack of tools, lack of funding for knowledge translation, little involvement of key stakeholders in the research process and competition between stakeholders.
"Do we have all the skills necessary, or the time even, to perhaps advise our partners how that's to have a policy impact ..."
" needs to do a lot more consultation with stakeholders from the start, so that consensus and coalitions supporting reform are established and gain momentum"
None of the respondents mentioned criteria regarding the type of knowledge or evidence needs to be translated into policy and practice, or when knowledge translation needs to be done.
Eighteen of 23 funding agencies describe some aspect of knowledge translation in their mandate (Table 3). However, the activities and definition of knowledge translation varied dramatically across different funding agencies, ranging from dissemination to brokering between researchers and decision-makers (Table 3).
"We're not an activist funding organization, per se. That's where the broker versus advocate role comes in."
Nine of 23 agencies described the focus of the knowledge translation activities as ensuring that funded research contributed to improving the health of their communities.
Budget and priority
Eight of 23 agencies ranked knowledge translation as a high priority. Seven of 23 agencies were able to report the percentage of their total budget spent on knowledge translation; all reported less than 20%. Three agencies reported that the knowledge translation budget would withstand cutbacks to the total budget.
Dissemination
One-third of agencies viewed dissemination as a shared responsibility between researcher and the funding agency. Others defined the main responsibility for dissemination as the role of researchers, funding agencies or partners. Dissemination activities were described as highly variable.
Most of the activities that agencies required, expected or encouraged by researchers were traditional within science communication such as producing a final report or journal publication. Thirteen of 23 agencies also required or encouraged researchers to partner with decision-makers and research users. Six agencies stated that researchers were encouraged to engage in pull activities that aim to increase the appetite for research by decision-makers. For example, Pan American Health Organization (PAHO) supported national research councils, including ministries of health.
Application process
At the time of application, 15 of 23 agencies described a requirement to partner with decision-makers, 12 of 23 agencies required researchers to state the policy relevance and significance of their research, and 11 of 23 agencies required researchers to define a knowledge translation audience (Table 4). Other activities described at the application stage were provision of a lay summary proposal, and a knowledge translation plan including dissemination, web development, publication and conferences (Table 5).
Agency initiatives
The agencies used five general strategies to support knowledge translation. These were classified as push, pull, linkage/exchange, communication and funding opportunities.
Funding mechanisms to promote knowledge translation included funding teams (including research users); funding conferences of researchers and research users; knowledge translation requests for applications; funding special centres and chairs for knowledge translation; and seeking commercialization opportunities (Table 6).
Twenty-two of 23 agencies described active involvement in communication activities such as communication to different audiences through web sites and paper journals (Table 6). These included development of audience-tailored web pages such as the South Africa Medical Research Council's AfroAIDS web site (available at: http://www.AfroAIDSinfo.org), lay summaries and use of media.
Linkage/exchange activities were described by 22 of 23 agencies, and included activities such as consulting stakeholders to set the research agenda, creating networks and programmes for decision-makers (Table 6). For example, the Indian Council of Medical Research funded partnerships with the private sector to improve access and availability of drugs for diseases of poverty, such as typhoid and measles vaccines.21
Half the agencies described some type of pull activity to increase skills of policy-makers to use research or increase their involvement in setting the research agenda, and fewer of these activities were described by each agency than the push and linkage/exchange types. These activities included tools development, programmes for decision-makers and workshops for decision-makers. For example, the Philippines Council for Health Research and Development described hosting research forums to expose decision-makers to research evaluation and critical appraisal.
The research team selected seven examples of innovative techniques ("gems") based on how they illustrate the diversity of ways in which funding agencies are engaging in knowledge translation (Table 7).
Equity
Nine agencies described poverty reduction or improved health equity as part of their main focus. Examples of equity-focused knowledge translation activities by funding agencies included: the WHO/TDR (Department of Research and Training in Tropical Diseases) programme to eliminate leprosy, the investment in schistosomiasis research in Brazil by FAPESP (Foundation for Research Support of the State of São Paulo), support of higher education for women and girls by USAID (United States Agency for International Development), and the destigmatization of groups at high-risk for HIV/AIDS sponsored by CIDA (Canadian International Development Agency).
Evaluation of agency activities
Thirteen agencies described evaluation tools to assess whether projects met their expectations. Eight agencies reported that they had an evaluation framework for knowledge translation activities. Tools used to evaluate the impact of knowledge translation activities were: (1) client/user surveys to assess how knowledge was used in practice and policy, and which products were most effective and useful; (2) visits to web sites; (3) number of telephone or e-mail queries on an information system; (4) requests for information from research users; and (5) outcome mapping.22
"There was a study..., [which showed that] only about 15% [of research funded by our agency] has been translated, meaning actually utilized into something commercialized, adopted really utilized."
Target audience
All funding agencies described several target audiences. The most commonly described target audience was decision-makers (16 agencies) and academics (12 agencies), followed by hospital managers (10 agencies), practitioners (10 agencies), other researchers (9 agencies), industry (9 agencies), researcher funders (8 agencies), general public (7 agencies), health-care professional organizations (7 agencies), media (6 agencies) and consumer organizations (3 agencies).
National versus international funding agencies
In this sample, the national agencies engaged in more knowledge translation activities than their international counterparts across all categories. For example, more national agencies required researchers to provide a knowledge translation plan (10/14 versus 3/9), identify a target audience (8/14 versus 3/9) and provided a budget for workshops (8/14 versus 1/9). More national agencies reported issuing requests for applications on knowledge translation using the media (13/14 versus 4/9) and stakeholder consultation (13/14 versus 6/9). The World Bank in the Philippines was a notable exception to other international funding agencies, as it had strong knowledge translation activities globally.
Discussion
This was a descriptive, exploratory study which identified substantial interest in knowledge translation of research results by both national and international funding agencies that support research in LMICs. We generated four hypotheses useful to studying the role of funding agencies in knowledge translation. First, national funding agencies in this sample demonstrated a greater commitment to knowledge translation activities than international funding agencies. Second, adoption of a systematic framework to knowledge translation might contribute to conceptual clarity in this field. Third, knowledge translation frameworks need to be modified to capture activities by funding agencies. Fourth, funding agencies are moving away from traditional methods of disseminating results and are being creative about reaching relevant audiences.
These findings suggest that national agencies may be more motivated to engage in knowledge translation activities than international funding agencies (with the exception of the World Bank in the Philippines). These findings lend credence to the perception that international funding agencies may not be well connected to realities on the ground at country-level. Furthermore, these findings support the focus on increasing funding for national health research within countries, as recommended by the Commission on Health Research for Development in 1990 (Karolinska Institute, Sweden). However, since international funding agencies still support over 90% of research in some low-income countries,16 their lack of focus on knowledge translation is worrisome. Encouragingly, there was interest in all international funding agencies to increase their knowledge translation activities in the next five years.
A common terminology for knowledge translation could be useful in better defining both existing and planned funding agency activities. We found different definitions and understanding of knowledge translation both within and between agencies (Table 3). The different terminologies reflect differences in the mandates of these organizations but also suggest a lack of conceptual clarity around knowledge translation.
We found a lack of consideration in determining which evidence required translation and the need for tailored approaches for different audiences. Despite the relatively incomplete evidence-base on the effectiveness of different knowledge translation strategies, there is evidence to support the use of audience-specific strategies (e.g. consumers, practitioners, policy-makers) to address audience-specific barriers and facilitators.2325 Furthermore, there are convincing arguments that knowledge transfer should be based on rigorous meta-analysis of systematic reviews based on all available studies rather than single studies, because systematic reviews increase confidence in results, reduce the chances of being misled and efficiently summarize all published literature.26 Adoption of a systematic framework to knowledge translation would contribute to conceptual clarity in this field. For example, the five step approach to knowledge transfer, described by Lavis, provides a framework to assess what should be transferred, to whom, by whom, how and with what effect.24
We found that the Lavis framework of push, pull and linkage/exchange was a useful tool to categorize knowledge translation activities. However, we found that these three categories alone did not capture all of the activities of funding agencies, therefore we added two codes for general activities by funding agencies: communication and funding opportunities. These five categories represented mutually exclusive codes that provided a useful basis for classifying activities. In our analysis of the discrepancies in coding between country teams and the Ottawa team, we found the greatest differences in interpretation between the push and communication categories. Our category of push was intended to capture activities that focused on researchers summarizing the actionable messages based on their research, going beyond traditional publications or reports to stating the policy relevance of their research findings.
We found several creative and innovative strategies such as the "gems" in Table 7. These creative strategies show that funding agencies are moving away from traditional methods of disseminating results.
Ability to evaluate the impact of knowledge translation strategies was lacking in all agencies. Lack of evaluation frameworks limit the ability to show whether knowledge translation efforts indeed enhance research-related policy, services (health and intersectoral) and societal impacts.27
Knowledge translation is a complex process which can enhance the health of disadvantaged populations, by improving access, diagnostic accuracy, compliance and adherence of effective services.3,13 We found a commitment to enhancing health of disadvantaged populations by one-third of funding agencies. We also found examples of knowledge translation activities that were focused on enhancing the health of the disadvantaged, such as the WHO/TDR programme to eliminate leprosy. Increased focus is needed to ensure that knowledge translation activities benefit the most disadvantaged populations.
An increasing number of organizations internationally are dedicated to knowledge translation. The activities of these organizations were not captured by our study, such as the WHO/PAHO EVIPNet), the Overseas Development Institute's RAPID programme and the Getting Research into Policy and Practice (GRIPP) initiative. These international initiatives represent an exciting opportunity to explore the effectiveness of different knowledge translation strategies.
Our results may overestimate the amount of knowledge translation activities since any activity (no matter how small) was scored as a "yes". We only interviewed three people from each agency so we may not have captured all knowledge translation activities. However, we tried to ensure interviewees represented a senior policy-maker, someone responsible for knowledge translation and a project officer. Three funding agencies interviewed for this study did not consider knowledge translation a main part of their mandate. This data was collected between September 2003 and September 2004, before the Ministerial Summit on Health Research convened by WHO in Mexico. Advocacy for knowledge translation has increased since the Summit, but it remains to be seen if funding agencies have actually shifted significant resources to this important area. This study provides a useful scan of the activities of these 23 agencies and the types of activities in which they are engaging.
Because this is a qualitative research study that used a judgement sample, we focused less on external validity and more on maximizing internal validity. Therefore, these results apply to the sample of funding agencies selected and included in this study and are not intended to be generalized to other funding agencies.
Conclusion
Previous research on knowledge translation has mostly ignored the role of funding agencies. This descriptive study shows an encouraging support for knowledge translation by national funding agencies, with a lag in support from international funding agencies. Funding agencies need to agree on a common terminology, consider the need for approaches tailored to specific audiences and identify their niche roles in knowledge translation, which may differ according to their defined mandates. Funding agencies might consider their role as knowledge brokers, by fostering and encouraging interactions between researchers and relevant stakeholders. As knowledge brokers, funding agencies could promote research syntheses and a focus on health equity. There is an urgent need to evaluate these funding agency knowledge translation activities to learn what works, why and in what context, in order to better justify spending on knowledge translation and to improve performance. 
Funding: Peter Tugwell is supported by a Canada Research Chair in Health Equity. Jeremy Grimshaw is supported by a Canada Research Chair in Knowledge Translation. Visanu Thamlikitkul is supported by The Thailand Research Fund. Vivian Welch is supported by a Canadian Institutes of Health Research doctoral scholarship.
Competing interests: None declared.
References
1. Neufeld V, Johnson N, eds. Forging links for health research: perspectives from the Council on Health Research for Development. International Development and Research Council; 2001.
2. Resolution WHA 58.34. Ministerial summit on health research. In: Fifty-eighth World Health Assembly, Geneva, 25 May 2005. Available from: http://www.who.int/rpc/meetings/58th_WHA_resolution.pdf [accessed on 19 February 2008] .
3. Bridging the "know-do" gap meeting on knowledge translation in global health, 10-12 October 2005. Geneva: WHO; 2005. Available from: http://www.who.int/kms/WHO_EIP_KMS_2006_2.pdf [accessed on 19 February 2008] .
4. Siddiqi K, Newell J, Robinson M. Getting evidence into practice: what works in developing countries? Int J Qual Health Care 2005;17:447-54. PMID:15872024 doi:10.1093/intqhc/mzi051
5. Translating research into practice (TRIP)-II. United States of America: Agency for Health Research and Quality; 2001.
6. Court J, Young J. Bridging research and policy in international development: context, evidence and links. In: Stone D, Maxwell S, eds. Global knowledge networks and international development. Routledge; 2004. Available from: http://www.odi.org.uk/Rapid//Publications/Documents/ODI_synthesis_TKN.pdf [accessed on 19 February 2008] .
7. Huskins WC, Ba-Thike K, Festin MR, Limpongsanurak S, Lumbiganon P, Peedicayil A, et al. An international survey of practice variation in the use of antibiotic prophylaxis in cesarian section. Int J Gynaecol Obstet 2001; 73:141-5. PMID:11336733 doi:10.1016/S0020-7292(01)00365-4
8. Heller RF, O'Connell RL, Lim LL, Atallah A, Lanas F, Joshi P, et al. Variation in stated management of acute myocardial infarction in five countries. Int J Cardiol 1999;68:63-7. PMID:10077402 doi:10.1016/S0167-5273(98)00343-X
9. Page J, Heller RF, Kinlay S, Lim LL, Qian W, Suping Z, et al. Where do developing world clinicians obtain evidence for practice: a case study on pneumonia. J Clin Epid 2000;53:669-75. doi:10.1016/S0895-4356(99)00231-0
10. Xu Q, Smith H, Li Z, Ji L, Garner P. Evidence-based obstetrics in four hospitals in China: an observational study to explore clinical practice, women's preferences and provider's view. BMC Pregnancy and Childbirth 2001;1:1. Available from: www.biomedcentral.com/content/pdf/1471-2393-1-1.pdf [accessed on 19 February 2008] .
11. Poverty and health: individual country reports. World Bank; 2000. Available from: http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTHEALTHNUTRITIONANDPOPULATION/EXTPAH/0,, contentMDK:20216965~menuPK:400482~pagePK:148956~piPK:216618~theSitePK:400476,00.html [accessed on 19 February 2008] .
12. Gwatkin DR. How well do health programmes reach the poor? Lancet 2003;361:540-1. PMID:12598134 doi:10.1016/S0140-6736(03)12558-5
13. Victora CG, Huicho L, Amaral JJ, Armstrong-Schellenberg J, Manzi F, Mason E, et al. Are health interventions implemented where they are most needed? District uptake of the integrated management of childhood illness strategy in Brazil, Peru and the United Republic of Tanzania. Bull World Health Organ 2006;84:792-801. PMID:17128359 doi:10.2471/BLT.06.030502
14. Tugwell P, de Savigny D, Hawker G, Robinson V. Equity-effectiveness loop: working against the odds: the application of clinical epidemiologic methods to health equity. BMJ 2005;332:358-61. doi:10.1136/bmj.332.7537.358
15. Morse J, Field P. Qualitative research methods for health professionals. 2nd edn. Thousand Oaks: Sage; 1995.
16. Lavis JN, Lomas J, Hamid M, Sewankambo NK. Assessing country-level efforts to link research to action. Bull World Health Organ 2006;84:620-8. PMID:16917649 doi:10.2471/BLT.06.030312
17. Tugwell P, Sitthi-Amorn C, Hatcher-Roberts J, Neufeld V, Makara P, Munoz F, et al. Health research profile to assess the capacity of low and middle income countries for equity-oriented research. BMC Public Health 2006;6:151. PMID:16768792 doi:10.1186/1471-2458-6-151
18. Patton MQ. Qualitative evaluation and research methods, 2nd edn. Newbury Park: Sage Publications; 1990.
19. Marshall C, Rossman G. Designing qualitative research. Newbury Park: Sage; 1989.
20. Crabtree B, Miller W. Doing qualitative research. Newbury Park: Sage; 1992.
21. Ganguly NK, Kant L. Closing gaps to achieve the MDGs: roles a medical research council can play: the Indian experience. In: Matlin S, ed. Global Forum Update on research for health 2005: Health research to achieve the Millennium Development Goals. London; 2005. pp. 60-63. Available from: http://www.globalforumhealth.org/filesupld/global_update1/GlobalUpdate1.pdf [accessed on 19 February 2008] .
22. International Development Research Centre (IDRC) Outcome Mapping. Available from: http://www.idrc.ca/en/ev-26586-201-1-DO_TOPIC.html [accessed on 19 February 2008] .
23. Tugwell P, Robinson V, Grimshaw J, Santesso N. Systematic reviews and knowledge translation. Bull World Health Organ 2006;84:643-51. PMID:16917652 doi:10.2471/BLT.05.026658
24. Lavis JN, Robertson D, Woodside J, McLeod CB, Abelson J. How can research organizations more effectively transfer research knowledge to decisionmakers? Milbank Q 2003;81:221-48. PMID:12841049 doi:10.1111/1468-0009.t01-1-00052
25. Santesso N, Tugwell P. Knowledge translation in developing countries. J Contin Educ Health Prof 2006;26:87-96. PMID:16557514 doi:10.1002/chp.55
26. Lavis JN, Davies HTO, Gruen RL. Working within and beyond the Cochrane Collaboration to make systematic reviews more useful to healthcare managers and policy makers. Healthcare Policy 2006;1(2):21-33.
27. Kuruvilla S, Mays N, Pleasant A, Walt G. Describing the impact of health research: a Research Impact Framework. BMC Health Serv Res 2006;6:134. PMID:17049092 doi:10.1186/1472-6963-6-134
(Submitted: 17 January 2007 Revised version received: 21 December 2007 Accepted: 2 January 2008 Published online: 1 April 2008)
1 Correspondence to Peter Tugwell (e-mail: elacasse@uottawa.ca).
doi:10.2471/BLT.07.040386